AI Under the Microscope: From Legal Battles and Security Risks to Breakthroughs and Creative Clash

1/

📰 OpenAI sued for defamation after ChatGPT fabricates legal accusations against radio host

🔥Strap in fellow Edgelords. This is bound to be the first of many legal cases surrounding GPT and its hallucinations. A radio host in Georgia is suing OpenAI after ChatGPT stated he had been accused of defrauding and embezzling funds from a non-profit organization. What’s at stake is establishing legal precedent to hold companies responsible and whether the case has substantial merit.

⚖️Why this matters: This is a real test for LLMs as a whole. There are no current legal presidents for these situations, so any court decisions will pave the way (or not) for these cases to continue. No doubt OpenAI has foreseen this situation happening, so it will be interesting to hear their defense. We’ve seen a lot of back and forth between Brands and the extent of liability when it comes to social media platforms and the content they host. So far, the courts have been very generous in favor of the platforms. Does a product or company have immunity based on their disclaimer? Are the courts even prepared to handle this kind of discussion? It will be interesting to see this play out, especially if a human can show direct harm to their very human life. Stay tuned to see how the value of a human life plays out against society altering software.

2/

📰AI Browser Extensions Are a Security Nightmare

🔥The first one to the finish line wins, right? Not when your security is on the line. No doubt you’ve all seen the absurdly vast list of AI extensions set to enhance your online browsing experience. There are 3 main points in this article, and well worth the 10 min read for your personal edification. From malware in disguise to data leaks, these handy tools are Pandora’s boxes waiting to be cracked open.

🦖Why this matters: Ian Malcom said it best in the dino park. We’re so busy thinking about if we could, we never thought about if we should. The race to market share leaves a trail of broken security in it’s wake. If you’re going to install sh*t on your browser, at least know the risks. Stay vigilant, educate yourself and your workforce, and don’t treat these extensions like innocent pests.

3/

📰New model offers a way to speed up drug discovery

🔥Using AI to improve human life is where it’s at. Researchers are applying the vast computing power and a new model, ConPLex (link), to match target proteins with potential drug molecules without having to perform the computationally intensive step of calculating the molecules’ structures. Hold onto your lab coats nerds. At it’s core, this system aims to remove the drudgery and time-sink that comes with medical research. Yay?

🧬Why this matters: You don’t have to be a computational scientist to appreciate the potential of this study. At its core, this endeavor aims to speed up the process of determining HOW different medications can interact and affect disease and conditions. Think of it as the 10X multiplier for meds. Which, as long as we can keep BigPharma’s money grubbing hands off of it, could lead to faster AND cheaper treatments.

4/

📰 MIT researchers make language models scalable self-learners

🔥MIT is killing it today. Working on “textual entailment”, developers are working to help LLMs more effectively understand human language and to create smaller models that work as well as the larger ones. Bonus: they’re also working on self-training without human supervision.

🤖Why this matters: Why indeed. Continual improvement in this realm has the potential to refine the machine learning landscape, creating models that better understand human language, are more scalable, trustworthy, and cost-effective. All good things. The self-training part is a bit troubling, however. As we’ve seen from Alice and Bob, Facebook’s negotiation bots back in 2017, AI without deliberate prioritization can and will find their own efficiencies for completing a task.

5/

📰 How AI art killed an indie book cover contest

🔥Artists find themselves in an awkward middle ground, fighting for relevance and recognition in a deluge of AI-aided (or outright created) competition. It’s a good read, complete with bad actors who take the money and run, well meaning humans struggling to accommodate “both sides” and an open door for philosophical discussions about the future of our own value.

😓Why this matters: he creative world will always be the first forced to their knees. In our modern society, the almighty dollar reigns supreme and the sanctity of human creativity is under siege. Artists – those dedicated visionaries in art, music, and writing – find themselves on the front line, as AI technology offers a cheaper and novel alternative to human-made art. The ramifications are more than monetary; it’s a crisis of identity and worth as life-long craftspeople are elbowed out by algorithms. These ripples are destined to spread. All professionals, not just artists, heed this warning – ‘efficiency’ may soon become the cloak that conceals the replacement of human touch in myriad fields. We’re going to have very real, very messy conversations about this one fellow humans. A warning to the non-creatives out there: This stuff is coming for you as well. Your replacement will simply be masked by “operational efficiencies”.

Thus brings our week to a close. Thank you to everyone who has joined up to follow along on this ride. If you’re so inclined, send us some feedback on how we’re doing. Be kind to each other this weekend and remember: we’re only human.