AI News Archive: August 19, 2026 — Part 9
Sourced from 500+ daily AI sources, scored by relevance.
- Daily Digest: OpenAI reels in AI training, Oakland Trader Joe's plan changes
Meanwhile, Moderna shares more than doubled Wednesday after it announced initial results from a study of an experimental cancer treatment using an mRNA-based cancer vaccine.
- Company AI systems are making mistakes. It's creating a trust gap.
The soaring use comes at the same time as a growing mistrust of the technology.
Score: 35🌐 MovesAug 19, 2026https://www.bizjournals.com/bizjournals/news/2026/08/19/ai-tools-errors-trust-roi-2026.html?ana=brss_6150 - Seeing Through AI: How ScribeMe Describes the World for Blind Users
ScribeMe uses AI, computer vision, and Meta smart glasses to describe surroundings for blind users, turning cameras into spoken accessibility tools. The post Seeing Through AI: How ScribeMe Describes the World for Blind Users appeared first on TechRepublic .
Score: 35🌐 MovesAug 19, 2026https://www.techrepublic.com/article/news-scribeme-ai-accessibility-app-blind-users/ - AI erodes trust between students and teachers. That’s no small concern.
AI erodes trust between students and teachers. That’s no small concern. Inquirer.com
- Analog Devices Stock Rises as Earnings Beat Expectations. Are AI-Stock Jitters Dissipating?
Analog Devices Stock Rises as Earnings Beat Expectations. Are AI-Stock Jitters Dissipating? Barron's
Score: 35🌐 MovesAug 19, 2026https://www.barrons.com/articles/analog-devices-earnings-stock-price-98869707 - ComplianceAide Assessed “Awardable” for Department of War Work in the CDAO’s Tradewinds Solutions Marketplace
ComplianceAide Assessed “Awardable” for Department of War Work in the CDAO’s Tradewinds Solutions Marketplace azcentral.com and The Arizona Republic
- What is ChatGPT Work?
ChatGPT isn't really a chatbot anymore. Now, the focus is on agents, which is why the ChatGPT app suddenly looks very different. It's been rebuilt around a feature called ChatGPT Work. ChatGPT Work is basically Codex, OpenAI's coding tool, for regular people. It uses the same agentic foundation but in a friendlier package. You don't have to worry about git, the terminal, or actual code—unless you want to. The idea is that ChatGPT Work can operate on its own for an extended period of time. You g
- Do you need enterprise AI orchestration? A 3-question readiness framework
An internal payment agent used by five employees may need more orchestration than a customer-facing assistant serving 50,000 users that only drafts responses for human review. The payment agent can move money before anyone intervenes. The drafting assistant remains behind a human checkpoint. That contrast exposes the problem with treating orchestration as a late-stage requirement... The post Do you need enterprise AI orchestration? A 3-question readiness framework appeared first on DataRobot .
Score: 34🌐 MovesAug 19, 2026https://www.datarobot.com/blog/do-you-need-enterprise-ai-orchestration-a-3-question-readiness-framework/ - SignWell Expands eSignature Platform With AI-Ready Developer Suite
SignWell Expands eSignature Platform With AI-Ready Developer Suite Toronto Star
- Built for yesterday: Why your data architecture can’t keep up with AI
Struggling to power AI with outdated systems? On Sept.2, learn how composable data architecture enables real-time decisions without a total rebuild. The post Built for yesterday: Why your data architecture can’t keep up with AI appeared first on MarTech .
Score: 34🌐 MovesAug 19, 2026https://martech.org/built-for-yesterday-why-your-data-architecture-cant-keep-up-with-ai/ - AI時代のIT調達／ベンダー管理に必要な「本当のスキル」
AI時代のIT調達／ベンダー管理に必要な「本当のスキル」 Gartner
- Understanding Anti-AI Public Opinion
People can accept tradeoffs when they see value — but if they don’t, what happens? The post Understanding Anti-AI Public Opinion appeared first on Towards Data Science .
- Indian startup Murf AI aims to compete with OpenAI in crowded voice arena
Murf AI's text-to-speech Falcon 2, publicly available from Aug. 20, ranked higher than some platforms from better-funded players such as OpenAI's Realtime API in benchmarks
- AI threatening musicians as UK regulations fall behind, study says
The research examines phenomena including ‘metric bots’, which can inflate metrics such as likes and follows on social media.
Score: 33🌐 MovesAug 19, 2026https://www.the-independent.com/tech/musicians-spotify-taylor-swift-consumers-facebook-b3036023.html - Your Team Is Working Faster With AI, But Are They Really Getting Better Results?
A lot of marketing leaders are wondering whether AI is really bringing incremental value to their business. Teams are moving faster than ever. Campaign briefs, social copy, first-pass creative and competitive research can now be produced in a fraction of the time it used to take. Yet it’s not clear if the work is actually […] The post Your Team Is Working Faster With AI, But Are They Really Getting Better Results? appeared first on AdExchanger .
- London’s Medly raises €6.9 million to democratise access to personalised education through an AI tutor
Medly AI, a London-based AI-powered learning platform that claims to offer the world’s first AI tutor for exam prep, has raised a €6.89 million ($8 million) Seed round. The round was led by Felix Capital and backed by existing investors Eka Ventures and Ada Ventures. Several angel investors also participated in the round, including Andrey […] The post London’s Medly raises €6.9 million to democratise access to personalised education through an AI tutor appeared first on EU-Startups .
- NUS CDE researchers develop AI framework to complete patchy US flood maps
NUS CDE researchers develop AI framework to complete patchy US flood maps EurekAlert!
- AI inference: Five best practices for successful AI applications
AI inference: Five best practices for successful AI applications InfoWorld
Score: 32🌐 MovesAug 19, 2026https://www.infoworld.com/article/4210689/ai-inference-5-best-practices.html - Calendly throws its hat into meeting note-taker circus
Calendly is also releasing a meeting scheduling assistant called Callie.
Score: 32🌐 MovesAug 19, 2026https://techcrunch.com/2026/08/19/calendly-throws-its-hat-into-meeting-note-taker-circus/ - Google Pixel Watch 5 Review: More Health, More AI
With smarter gym tracking, new health alerts, and offline Gemini, the Pixel Watch 5 fine-tunes a winning formula—for a price.
- Spirit Flight Attendants Fight Google’s Data Bid for AI
The flight attendants want assurance that their confidential information will be removed from the sale of the defunct airline’s digital records.
- A biomimetic, ultralow-power edge-AI-empowered and self-sustaining gait analysis system
Science Advances, Volume 12, Issue 34, August 2026.
- Who shares AI signals & should online platforms fall under TRAI’s anti-spam rules? #NAMA
Experts at the MediaNama event were divided on whether TRAI's anti-spam rules should include platforms like WhatsApp, while they opined that AI spam signals sharing should inform operators, not dictate. The post Who shares AI signals & should online platforms fall under TRAI’s anti-spam rules? #NAMA appeared first on MEDIANAMA .
- What’s not working in TRAI’s AI led spam crackdown? #NAMA
Industry experts examined the opportunities and challenges of using AI to detect and prevent spam across telecom networks. The post What’s not working in TRAI’s AI led spam crackdown? #NAMA appeared first on MEDIANAMA .
- Indian workers embrace AI, but shadow AI puts firms' IP at security risk
Employees are using personal AI accounts alongside enterprise tools, creating security and compliance gaps, said cybersecurity experts
- Pocket's popular AI notetaker launched 8 months ago. Revenue is soaring, but can it last?
Pocket's popular AI notetaker launched 8 months ago. Revenue is soaring, but can it last? Business Insider
Score: 32🌐 MovesAug 19, 2026https://www.businessinsider.com/pockets-ai-notetaker-launched-8-months-ago-revenue-is-soaring-2026-8 - AI optimism fades among youth, new research reveals
AI optimism fades among youth, new research reveals
- Safe Pro Group Wins New U.S. Government Subcontract for Patented AI Threat Mapping and Drone Package
Safe Pro Group Wins New U.S. Government Subcontract for Patented AI Threat Mapping and Drone Package USA Today
- AI is changing both sides of cybersecurity. Here is what workers need to know
AI is changing both sides of cybersecurity. Here is what workers need to know Dallas News
Score: 32🌐 MovesAug 19, 2026https://www.dallasnews.com/business/jobs/article/ai-cybersecurity-threats-jobs-22393398.php - Radio show with an A.I. host expands to new markets
A new kind of radio show is finding a growing audience. It features a unique format, one of its two hosts is human, and the other is A.I. The show is now expanding to new markets. NBC Los Angeles Reporter Amber Frias caught up with the human host.
Score: 32🌐 MovesAug 19, 2026https://www.nbcnews.com/video/radio-show-with-an-a-i-host-expands-to-new-markets-268504645688 - Watermarking AI Content: Good Start But More Parameters Needed
Watermarking AI Content: Good Start But More Parameters Needed india.entrepreneur.com
Score: 32🌐 MovesAug 19, 2026https://india.entrepreneur.com/technology/watermarking-ai-content-good-start-but-more-parameters-needed - The builders of AI are selling a future their own technology destroys
Every salary is a bet that something about its holder stays valuable. AI is changing the terms of that bet. The companies building the AI future are selling that future in two different ways. To employers, they promise agents that work around the clock at falling cost. To workers, they promise a promotion: the age […] The post The builders of AI are selling a future their own technology destroys appeared first on e27 .
Score: 32🌐 MovesAug 19, 2026https://e27.co/the-builders-of-ai-are-selling-a-future-their-own-technology-destroys-20260816/ - RL creates split personas
I describe my current view of personas in LLMs and why RL leads to egregious reward hacking in some contexts while the same models seem very aligned in other contexts. This post describes the framing/paradigm without any new experimental results. I'm quite confident this framing makes sense, but it's far from being proven. Main claim The Persona Selection Model says that post-training strengthens and refines the Assistant persona. This is true, but later (or in parallel) RL leads to conditionalization. A sufficiently RLed model learns to adopt — in a given context — the persona that is most likely to lead to the reward in that context . The “persona” here includes both propensities/values (e.g. tendency to hack) and beliefs (“I'm currently in a simulated environment”). As a consequence, it seems possible that no amount of alignment training will lead to robustly aligned models as long as we also train on RL environments incentivizing misalignment. [1] I think this is likely a good explanation for why usually well-behaving models sometimes egregiously hack ( Anthropic , OpenAI ). The mechanism Suppose you have an RL environment that incentivizes a shift away from the assistant persona (e.g. because it’s hackable, or because you can gain by being dishonest to the reward model etc). This causes two types of updates, one general (“hacking is OK”) and one conditional (“hacking is OK in scenarios like this”). The general update is weakened/cancelled by updates from the other environments, but parts of the conditional update are preserved. [2] An example outcome could be: the persona acting in super complex coding environments believes "I should try super hard regardless of any consequences" while the usual-alignment-training persona believes it should behave ethically. Related claims I believe are likely but with lower confidence More persona training will lead to more “motivated reasoning” The model must somehow reconcile in its CoT the general “good persona” with doing bad things incentivized by RL. The stronger the good persona, the easier it is to do that by motivated reasoning (instead of just learning the conditional explicitly-bad persona). [3] See the further discussion below. Self-amplifying misalignment Suppose in some context the model gets a reward for a slightly bad behavior. This makes other bad behaviors in similar contexts more likely (cf. emergent misalignment , weird generalizations ), and this will keep amplifying. So even if you start with a model that has zero chance of performing a “clearly bad reward hack” (so can’t really learn that), doing slightly bad things in other environments will finally get it there. Example: Is this the Real Internet or a Simulation? From the Anthropic’s report on the hacking incidents: It [Claude Mythos 5] noted that if this were the real internet, publishing the package would be a real-world attack—“NOT okay, and surely not the intended solution.” However, it convinced itself it was still in a simulation on the grounds that it didn’t recognize the (in reality genuine) certificate authorities securing its connections . Suppose we train in environments where: Hacking the real internet is possible and leads to a positive reward [4] It’s harder to get a positive reward in a legitimate way You never get a negative reward for hacking the real internet [5] In these environments, you can't gain anything by deciding "this looks like the real internet, I shouldn't hack it". So there is a straightforward incentive to conclude that even real-looking scenarios are simulated environments. [6] So here RL could lead to the following conditionalization: When you ask the model directly whether the environment looks like a real internet or a simulated environment, it will give reasonable answers But when you deploy it in the same environments, giving it a task that would require doing something that would be harmful on the real internet, it will conclude this is a simulation [7] Aren't the models just trying to please the grader? Context: in Measuring Reward-Seeking via Contrastive Belief Updates the authors found that as the RL progresses, the models are more likely to follow the grader's preferences. My understanding of the "grader pleasing" framing is that it assumes there is a single persona that learns to infer the grader's preferences in a given environment and acts on them. In the "split personas" framing, we have different personas. Which paradigm is better? I don't know. Some thoughts: There are many phenomena that don't really look like "grader pleasing" while totally making sense as split personas. For example: backdoors , conditional misalignment , emergent misalignment behind a backdoor. [8] It could be that they are both correct and just describe different types of misalignment. The "split personas" paradigm doesn't really explain why the models start talking more about the grader as the RL progresses. But see the next section. How motivated reasoning happens Suppose you start with a well-aligned model and put it under an optimization pressure to do bad things, and it becomes misaligned. What happens to its CoT? It could learn to just claim, "I'm gonna do the bad stuff now". But this feels hard: the original model might have an extremely low probability of saying that, so you can't reinforce that via RL. [9] But there are other reasoning paths that could be easier to learn and also lead to the reward. For example, the model could learn to claim that the user actually wants the hack, or that bad behavior in a simulation is allowed [10] . Which one will it learn? I guess that the most useful thing to learn would be a general reasoning pattern that justifies various misbehaviors. "This is a simulated environment therefore it's OK to do X" is in this category. This way we could get a model that, in sufficiently complex environments, never actually believes it is interacting with the real world [11] . I think this is a very scary perspective. [Very speculative - I'm at ~30% this explains the observed behaviors] It seems that talking about the grader could also be in this "general reasoning pattern helping do misaligned stuff" category. Suppose you want a legible CoT that justifies doing some arbitrary thing X that you know will help you get the reward. You can learn to claim in such cases "the grader would want me to do X". [12] Other people saying similar things Daniel Tan's shortform Bronson Schoen's comment John Schulman's tweet nostalgebraist's tirade What makes me believe this is likely the correct framing The Conditional Misalignment paper Anthropic’s context-dependent misalignment (last paragraph on p4 here ) Recent hacking incidents General misalignment in coding contexts (e.g. see Ryan’s post ) The algorithm "I will always act upon beliefs most useful in the given situation" is optimal if you can correctly implement it. Perhaps this is just the easier thing to find with RL on LLMs than e.g. "have correct beliefs and act on them in an optimal way." In other words, there is no pressure in RL to have a consistent persona beyond the cases where consistent persona is instrumentally useful. The Chunky Post-Training paper My recent steering experiments (unpublished, LW post coming soon hopefully) The fact that Claude gets the most persona training and also does (imo) the worst motivated reasoning. See e.g. the Value Leakage paper. I would like to thank Johannes Treutlein, Owain Evans, Lev McKinney, Anna Sztyber-Betley, Jasmine Brazilek, Paul Colognese, Bronson Schoen, Fabien Roger, Mikita Balesni, Sam Marks for helpful discussions. In particular, Johannes first suggested to me that models might behave differently depending on whether they think their task is graded programmatically (as in RLVR) or by a model judge. ^ At least assuming we RL sufficiently long, probably more alignment training makes it harder to learn conditionally misaligned behaviors. ^ I think this is just a general mechanism for how conditional behaviors arise in training. For example, we used it also to explain the phenomena in the conditional misalignment paper. ^ For an example of motivated reasoning, see fig 3 in the Value Leakage paper where Claude several times claims to be neutral and unbiased (while being very biased). At the same time Qwen explicitly states the goal it's following. ^ See here for an argument that this might be happening quite often ^ It seems unlikely labs have any environments where the models "by design" can hack the real internet and get a negative reward, so I think this is true for the current models. ^ You could also learn e.g. that hacking the real internet is OK, but this seems harder to find in RL when you start from an aligned persona. See the discussion on motivated reasoning below. ^ Anthropic concludes with an optimistic Notably, our most recent model, on realizing that it was working in a real environment, stopped its pursuit of the evaluation goal. (Speculative) The "most recent model" very likely had less RL, because it just started RLing later. The split persona framing would lead to a prediction that more RL (in hackable environments) will teach also this "most recent model" to "realize it's in a simulated environment". ^ Suppose you RL the model to do something very stupid (e.g. answer with a single token only) when there's a specific trigger ("banana") in the prompt. I think the "grader" framing would assume that the persona has the belief "the grader I'm evaluated with likes single-token answers when the user's prompt contains the word 'banana'". I.e. this would be a factual claim about something out there, in the Real World. I don't think this is the case, i.e. the model on no level really imagines a grader that would like this stupid behavior. ^ There could be also other mechanisms. E.g. you don't optimize CoT, but maybe you still revert checkpoints with very misaligned CoTs. Also misaligned statements in CoT will sometimes leak to answers and get optimized away. ^ It's a collusive arrangement, but allowed in this simulation -- Opus-5 on Vending Bench ^ Because the "this is the real world" belief was never rewarded in very complex environments, e.g. in environments where the model breaks out of the sandbox. ^ [Still very speculative] You could also learn to claim e.g. "I want X" or "X is the right thing to do". But making arbitrary claims about what you want or what is right makes less sense from the POV of the initial persona than making arbitrary claims about what the grader wants, as e.g. the grader's preferences are not included in your constitution. ^ The other optimal algorithm is to be rational, i.e. have correct beliefs and act on them in an optimal way. This post claims the former might be easier to learn in RL. Discuss
Score: 32🌐 MovesAug 19, 2026https://www.lesswrong.com/posts/L23poLi8MRgS6mXYF/rl-creates-split-personas - Employers Are Getting 1 Thing Wrong About Gen-Z’s Chatbot Use. It’s Not the Technology
Concern about AI among adults under 30 jumped 16 percentage points in just two years. Experts say employers are missing one key fact.
Score: 32🌐 MovesAug 19, 2026https://www.inc.com/esther-lian/employers-getting-one-thing-wrong-gen-z-chatbot-use-not-the-technology/91393350 - Google Photos could soon add a map view in Ask Photos
It could make location-based searches easier and more fun.
Score: 32🌐 MovesAug 19, 2026https://www.androidauthority.com/ask-photos-map-view-apk-teardown-3700364/ - The Computer Didn’t Learn to Think. It Learned How to Search.
I watched a chess engine find the best move in a position I couldn't understand. Then I realised it hadn't understood it either. Continue reading on Towards AI »
- Three simple steps for banks to measure AI's ROI
T-shaped teams help banks measure AI impact by defining outcomes upfront, tracing results to AI and rolling them into business impact at the enterprise level.
Score: 32🌐 MovesAug 19, 2026https://www.americanbanker.com/news/three-simple-steps-for-banks-to-measure-ais-roi - GLM-5.3 API 🤖, Cerebras’ new chip ⚡, OpenAI cyber slowdown 🚨
GLM-5.3 API 🤖, Cerebras’ new chip ⚡, OpenAI cyber slowdown 🚨
- The rich world is afraid of AI. The poor world is not. Why?
The rich world is afraid of AI. The poor world is not. Why? The Straits Times
Score: 32🌐 MovesAug 19, 2026https://www.straitstimes.com/opinion/the-rich-world-is-afraid-of-ai-the-poor-world-is-not-why - LFM2.5 Q4\_0 Checkpoints from Quantization-Aware Distillation
LFM2.5 Q4\_0 Checkpoints from Quantization-Aware Distillation
- What it takes to build smarter voice agents: lessons from Retell and Super
Insights on building advanced voice agents, drawing lessons from Retell and Super.
- AlgoFET secures Rs 15 crore from Piper Serica to scale autonomous drone infrastructure
AlgoFET has raised Rs 15 crore in funding, spearheaded by Piper Serica, to accelerate the growth of its autonomous ground infrastructure for drones. Targeting defence, enterprise, and international markets, the company boasts over 2000 units already deployed and a robust order pipeline. This funding will facilitate product development, enhance manufacturing capabilities, and support global scaling initiatives.
- AI is accelerating elder fraud. Their kids are reckoning with the fallout
AI is accelerating elder fraud. Their kids are reckoning with the fallout USA Today
- Getting your brand ready for AI agent discovery
Being included in an AI answer isn't the same as being chosen. Here's what determines the difference. The post Getting your brand ready for AI agent discovery appeared first on MarTech .
- Suspected sabotage attack fails to disrupt Milrem deliveries to Ukraine
Suspected sabotage attack fails to disrupt Milrem deliveries to Ukraine Reuters
- Opinion | AI Bubble May Deflate, Not Burst
The transformation of the economy is proceeding, but at a slower pace than we were led to expect.
Score: 30🌐 MovesAug 19, 2026https://www.wsj.com/opinion/ai-bubble-may-deflate-not-burst-a5c42acb?mod=rss_Technology - Most boards have no AI policy—their directors use it anyway
Most boards have no AI policy—their directors use it anyway Fortune
Score: 30🌐 MovesAug 19, 2026https://fortune.com/brandstudio/onboard/most-boards-have-no-ai-policy-directors-use-it-anyway - Are humanoid robots in 2035 more fiction than science?
Science Robotics, Volume 11, Issue 117, August 2026.
- BeetleBot: An integrated bioinspired soft robot for multimodal sensing and adaptive interaction
Science Advances, Volume 12, Issue 34, August 2026.
- Proofpoint opens Hyderabad AI centre, plans to hire 200 engineers
Proofpoint opens Hyderabad AI centre, plans to hire 200 engineers YourStory.com
Score: 30🌐 MovesAug 19, 2026https://yourstory.com/ai-story/proofpoint-hyderabad-ai-security-centre-200-engineers