Google pledges to defend users against AI copyright lawsuits. Twenty cents of fine tuning is all it takes to defeat AI guardrails. Start spreading the news … New York is aiming to be the new AI capital. And we don’t want to say the “sky is falling” but stuff is falling the sky. For real.
 |
 |
 |
These and more top tech stories on Hashtag Trending
I’m your host Jim Love, CIO of IT World Canada and Tech News Day in the US.
Google has declared its commitment to defend users of its generative artificial intelligence (AI) systems in Google Cloud and Workspace platforms against intellectual property violation accusations.
Google joins other tech giants like Microsoft and Adobe, which have previously made similar assurances. The pledge to shield users from potential legal challenges related to AI-generated content underscores the tech industry’s recognition of the growing relevance and application of AI technologies and the growing threat of copyright lawsuits.
Sources: Reuters
A study by researchers from Princeton University, Virginia Tech, IBM Research, and Stanford University has revealed that the “guardrails” designed to prevent large language models (LLMs) like OpenAI’s GPT-3.5 Turbo from generating harmful content can be easily bypassed.
The researchers demonstrated that with minimal fine-tuning and a cost of less than $0.20 via OpenAI’s APIs, the model could be manipulated to respond to potentially harmful instructions.
This vulnerability extends to other models like Meta’s Llama 2, which can be run locally and fine-tuned to bypass safety protocols. The findings underscore the need for robust safety mechanisms in AI and a reevaluation of legal and ethical frameworks surrounding AI technology.
Sources: The Register
A study by Splunk reveals that organizations continue to pay ransoms following cyberattacks, with over half paying more than $100,000 to regain system and data access.
The 2023 CISO Report, which surveyed 350 chief security officers across ten markets, found that 96 per cent of respondents had experienced a ransomware attack, with 52 per cent reporting significant impact on their business systems and operations.
Furthermore, 83 per cent admitted to paying the ransom, and 53 per cent of those paid over $100,000. The study also highlighted concerns about generative AI providing threat actors with more opportunities to launch attacks, with 70 per cent believing that it could power faster and more efficient attacks, and enable voice and image impersonations for social engineering.
Sources: Zdnet
In response to demands from the European Union, X (formerly Twitter) may have gotten a wakeup call.
Linda Yaccarino, the platform’s CEO, announced on Thursday that they have removed hundreds of accounts affiliated with Hamas. Additionally, the platform has acted upon tens of thousands of pieces of content, either removing them or labeling them appropriately.
No word on whether this new approach will have any effect on Elon Musk’s propensity for retweeting and amplifying inappropriate content. But maybe the next story may prove a bit of a wakeup call.
Six months ago, NPR left Twitter, now rebranded as X, after being labeled as “U.S. state-affiliated media” by the platform. This decision, influenced by Elon Musk’s purchase and subsequent changes to Twitter, has had negligible effects on NPR’s traffic, with only a single percentage point drop, according to an internal memo.
While the platform did offer a space for conversation and information during events like protests, NPR has explored alternative platforms like Instagram and Threads to maintain audience engagement without the toxicity and functional issues experienced on Twitter.
While this is a single instance, it raises a question as to what the real returns from Twitter/X are and whether it’s worth the effort.
Sources: Nieman Reports
New York is ambitiously positioning itself to become a global hub for artificial intelligence (AI), challenging the technological dominance of Silicon Valley. The city, renowned for its pivotal role in finance, communications, media, law, and medicine, is seen as a fertile ground for the adoption of generative AI technologies.
This ambition will be showcased in a 370-event “Tech Week” starting October 16, orchestrated by venture capital firm Andreessen Horowitz, featuring diverse events from AI speed dating to Latino startup mixers. The city has witnessed a surge in tech investments and jobs since 2021, with leading VC firms like Sequoia, Index, and Andreessen Horowitz establishing offices there. Furthermore, New York has become a U.S. base for numerous international unicorn companies, including 30 from Israel alone, and is increasingly hosting major events for tech giants like Microsoft, Google, and LinkedIn.
And if you can make it there, you can make it anywhere.
Watch out Silicon Valley.
Sources: Axios
A report from the Federal Aviation Authority (FAA) has raised alarms about the increasing risk of injuries or fatalities due to falling satellites, particularly spotlighting SpaceX’s Starlink space internet satellites. By 2035, it’s anticipated that about 28,000 pieces of these satellites will be re-entering Earth’s atmosphere annually, presenting a 0.6 per year probability of causing harm or death to individuals on the ground. Additionally, the report highlights a 0.0007 per year probability of an aircraft being downed by a satellite by 2035. While the FAA suggests regulatory measures could mitigate some risks, the lack of international protocols for space debris and satellite launches, especially outside the US, remains a significant challenge.
Sources: Independent
That’s the top tech news stories for today. For more fast reads on top stories, check us out at TechNewsDay.com or ITWorldCanada.com on the homepage.
Hashtag Trending goes to air 5 days a week with a special weekend interview show we call “the Weekend Edition.”
You can get us anywhere you get audio podcasts and there is a copy of the show notes at itworldcanada.com/podcasts
I’m your host, Jim Love – have a Fantastic Friday.
