Top Companies' AI Moves

Today's Stories

1. 1

   Smartphone AI detects eye diseases in real-world study without internet

   Researchers tested offline AI running on smartphones to detect diabetic retinopathy, glaucoma, and age-related macular degeneration (AMD) in a real-world setting. The AI operates without an internet connection, making diagnosis possible in settings where connectivity is limited. Eye disease detection often requires specialized equipment and trained professionals unavailable in many parts of the world. A smartphone-based tool that works offline could expand access to screening in underserved communities and potentially catch these conditions earlier, when treatment is more effective.

   The study demonstrates that offline AI can perform accurate diagnosis in practical conditions, which may influence how healthcare providers in low-resource settings approach eye disease screening.

2. 2

   MediaScience Study: AI-Generated Coca-Cola Ad Matches Original in Consumer Impact

   MediaScience conducted a study comparing an AI-generated Coca-Cola commercial with the original ad, finding that the two matched in consumer impact. The finding suggests AI-generated advertising creative can achieve similar effectiveness to human-produced content, which may reshape how brands approach commercial production and budgeting decisions.

   This result could influence advertiser confidence in AI creative tools, though broader industry adoption will depend on how consistently AI-generated ads perform across different brands and campaign types.

3. 3

   Philip Morris accused of flooding EU tobacco consultation with AI-generated responses

   Dutch broadcasters Pointer KRO-NCRV and NOS reported that Philip Morris International used its 'Your Voice, Your Choice' platform to direct users to a tool offering around 50 pre-defined responses opposing stricter tobacco regulations. AI detection analysis of over 65,000 responses from 15 EU countries found that 35 per cent were generated by AI, with 71 per cent of the 786 Dutch responses being almost certainly AI-generated. Over 80,000 responses were submitted to the European Commission's public consultation on revising tobacco legislation. The high volume of AI-assisted submissions raises concerns about the integrity of EU democratic consultations, which normally receive fewer than one thousand responses. Philip Morris stated users retained full control to edit responses, but critics, including France's National Committee Against Smoking, characterize the campaign as industry manipulation.

   The European Union is revising the Tobacco Products Directive and the Tobacco Advertising Directive to tighten rules on e-cigarettes, nicotine pouches, and heated tobacco. How the Commission weighs the AI-generated responses against substantive regulatory feedback will signal whether public consultations remain credible democratic instruments.

4. 4

   Tesla caps AI spending at $200/week, exempts Musk's xAI products

   Tesla imposed a $200-per-week spending limit on employee AI usage starting July 6, according to an internal memo, but excluded beta versions of xAI products from the cap. The move reverses course after six months of pushing engineers to use AI more aggressively, during which some software engineers were consuming thousands of dollars' worth of tokens each week. Tesla's entire valuation now rests on AI deployment across its Robotaxi network and Optimus humanoid robot, yet the company cannot manage its own internal token spending without capping it—raising questions about whether it can scale AI across its planned fleet. The exemption for xAI products steers employees toward Musk's own AI company rather than rivals like Claude, but Grok remains unpopular among Tesla staff despite the internal push.

   The carve-out reveals that Musk is using a spending policy to funnel employees toward in-house tools while his own engineers prefer competitors' products. This mirrors Uber's $1,500-per-month cap after burning through its 2026 AI budget by April, and similar cost-control moves at Meta, Amazon, and Walmart—all signaling that token-based billing is forcing corporate America to rein in AI spending.

5. 5

   OpenAI and Visa: AI now does the shopping for you

   OpenAI and Visa: AI now does the shopping for you

6. 6

   Microsoft launches $2.5B unit to embed AI engineers at customer sites

   Microsoft announced a new division called Frontier Company with $2.5B in funding, designed to place AI engineers directly inside customer organizations to help them adopt and integrate AI technologies. Rather than selling AI tools off-the-shelf, Microsoft is betting that embedding specialized engineers on-site will help enterprises overcome adoption barriers and realize AI value faster—a shift toward high-touch consulting alongside software products.

   The Frontier Company will operate as a distinct business unit within Microsoft, signaling the company views this embedded engineering model as a core long-term strategy, not a temporary service.