OpenAI’s new Codex system tries to answer that question by translating users’ natural language into code. Though this system could one day improve programming speed and efficiency, the current model lacks accuracy and is susceptible to algorithmic bias.
Winners of the company’s first-ever AI bias bounty program found Twitter’s image-cropping tool favors young people with thin figures and light, smooth skin. The bounty program could offer a model for other firms to increase transparency and build trust but won’t solve AI bias.
Pizza Hut looks outside of the box: The fast-food chain seeks to regain leadership in digital food delivery through AI and analytics.
Apple plans to introduce software to detect child abuse content locally on an iPhone or Mac. Privacy and security advocates say the plan could tarnish Apple’s reputation as the vanguard of consumer privacy and bulwark against government anti-encryption efforts.
Digital health solutions work better together: Innovaccer added healthcare AI firm Jvion’s prescriptive analytics platform to its suite of cloud-based digital health solutions—we unpack how smaller digital health cos are riding on the coattails of larger ones.
Twitter is offering researchers up $3,500 to uncover bias in its image cropping algorithm. Though bias bounty programs may reveal some instances of bias, they cannot hope to solve the broader issue of algorithmic bias entrenched within the bedrock of artificial intelligence.
The president’s pick for assistant secretary of state for business focuses on the impact of AI financial crises. The move is the latest attempt by the administration to work toward a unified national AI strategy amid rising global competition.
The brain-computer interface startup received $205 million in a recent funding round. Though BCIs are on the rise, interfaces that require invasive surgery like Neuralink may face steeper adoption hurdles.
Lawmakers try to chip away at Section 230: The proposed legislation would remove liability protection for companies that spread health misinformation via algorithms. But the difficulty of defining “misinformation” and determining which content was algorithmically amplified presents challenges to the bill’s adoption.
Amazon announced most of its Echo devices will support the cross-platform smart home standard— an essential endorsement for Matter’s success given Amazon’s outsize role in the smart speaker market.
Facebook’s CEO believes its hefty investment in AR and VR could make it a powerful player in the next stage of the internet. The proclamation comes amid rampant regulatory scrutiny and slowing user growth among Facebook’s core products.
New safety issues involving Tesla’s Full Self-Driving feature call into question the company’s choice to test on public roads. Though rapid real-world testing could advance the tech, it may also amplify consumers’ No. 1 concern about AVs: safety.
A pair of government watchdog reports into federal use of facial recognition has reinvigorated bipartisan calls for AI regulation generally. Algorithmic amplification of online content could present an opportunity for consensus for both parties, albeit from different angles.