Daily AI NEWS - July 28, 2024
Strikes, AI, and Privacy

Members of the union SAG-AFTRA have voted to go on strike against major video game studios. The performers are demanding additional A.I. protections, fearing their voices and likenesses could be replicated without consent. The strike is set to begin this Friday and is expected to impact many games.

Despite the strike, the highly anticipated Grand Theft Auto 6 will not face delays. Development on the game remains unaffected, offering some reassurance to fans. Meanwhile, the strike highlights ongoing concerns about the ethical implications of artificial intelligence in the entertainment industry.

OpenAI has launched SearchGPT, a new search engine prototype aiming to challenge Google's dominance. The tool promises fast and timely answers with clear and relevant sources.

Company CEO Sam Altman believes "there is room to make search much better than it is today," positioning SearchGPT to offer superior search experiences compared to existing platforms.

With the official waitlist now open, this move intensifies competition in AI-driven search, as Microsoft also vies with Google in this space.

Open AI models that allow developers to customize them with few restrictions are more likely to promote competition, FTC Chair Lina Khan said, weighing in on a key debate within the industry. From a report: "There's tremendous potential for open-weight models...

Elon Musk's X, formerly Twitter, has quietly begun using users’ tweets to train its AI, Grok, sparking outrage. The data collection is set to opt-out by default rather than opt-in, without prior notice. Users can stop X from scraping their posts by changing their privacy settings.

Musk touts Grok’s real-time access to X as its key strength. This means it uses live user data to refine its generative AI capabilities, ensuring it stays ahead of the competition. However, the lack of transparency has frustrated many, especially with the sudden policy change.

With Grok's next-gen model, Grok 2, imminent, X aims to compete more aggressively in the AI field. Despite Musk's claims of it being 'the most powerful' AI, the methods of data collection by default—without clear user consent—remain controversial, raising concerns about privacy and ethics.

Recent studies have raised alarms about the potential risks of AI "inbreeding". This issue arises when models like ChatGPT and Microsoft Copilot are recursively trained on their own data, creating self-perpetuating cycles. Researchers warn that this practice could lead to the eventual collapse of these AI tools.

The concern is that continuously using the same data can result in degraded performance over time. As these AI systems become more reliant on their own outputs, the quality and diversity of information they provide may suffer, impacting their long-term viability and accuracy.