Plus, AI Diagnoses Alzheimers Earlier
Ukraine’s military has been using AI technology to create drones that can operate in war conditions and are less vulnerable to electronic jamming, much to the worry of security experts about nefarious exploitation. In one test, a drone that was dropped connection could still blast its target due to new AI software. Over 200 Ukrainian companies are working with the country’s military to improve drones and increase production numbers. The drones are already impacting the battlefield, enabling Ukraine to destroy Moscow’s vehicles and damage the Crimean Bridge. These developments will also be a game-changer for states and forces wishing to have a technological edge in conflicts, making AI used in drones one of the most interesting sectors of the weaponization industry. In recent years Russia has begun testing its own drones as the selling of military drones proves an increasingly profitable business. However, drones should not be seen as a panacea for the Ukrainian side as their payload size and range are limited, and their chances of success may not extend past intelligence gathering. Owing to these drone improvements, there has been a rapid proliferation of people seeking to understand the basics of drone warfare.
The owner of G/O Media, which includes websites such as Gizmodo and Deadspin, recently published four articles generated entirely by AI. However, the AI-generated articles were filled with inaccuracies and failed to fulfill the promise of their headlines. Gizmodo’s union criticized the publication of AI-generated content as unethical, noting that it generates false information and plagiarizes work. Despite the objections, G/O Media defended its decision, viewing the trial as successful and indicating a desire to use AI more frequently. The motive behind using AI in media is to reduce labor costs and maximize profits drastically. However, this approach sacrifices the value proposition of providing truthful and insightful information to readers. In the case of G/O Media, the AI-generated articles lacked personal touches, used awkward phrasing, and contradicted their headlines. The media company’s focus on AI-generated content demonstrates a disregard for journalism and public knowledge. The use of AI as a means to cut costs and increase profitability can undermine the reputation and integrity of media brands.
China is rapidly developing its military AI capabilities. These advances could threaten the US, as China could use AI to develop more sophisticated weapons and surveillance systems. The US needs to take steps to counter China’s AI advances or risk falling behind in the military technology race.
An Indian news channel has introduced AI-generated news anchors to read the headlines during broadcasts. Odisha TV uses an AI anchor called Lisa, while Aaj Tak employs an AI anchor named Sana. The robotic anchors can read headlines in multiple languages and are being used to complement the work of human presenters. However, there has been a mixed reception to using the AI anchors. Supporters argue that they can provide news faster and in multiple languages during critical times such as elections. On the other hand, critics are concerned about AI replacing humans and the lack of human nuances in the presentation. There are also ethical concerns about the potential for racial and sexist discrimination when creating AI bots in the image of human beings. Both Odisha TV and Aaj Tak claim that the AI anchors are not meant to replace their human counterparts but rather to handle repetitive and mundane tasks.
Researchers have developed an AI tool that can help detect Alzheimer’s disease earlier. The tool analyzes brain scans and can identify patterns associated with the disease. This could help doctors diagnose Alzheimer’s disease earlier, improving treatment outcomes.
OpenAI has decided to shut down its AI classifier tool that aimed to identify AI-generated writing due to its low accuracy rate. The company plans to develop mechanisms for users to discern whether audio or visual content is AI-generated instead. OpenAI acknowledged that the classifier was not effective at catching AI-generated text and had the potential to produce false positives by tagging human-written text as AI-generated. The decision to close the tool came after concerns were raised about the impact of AI-generated text, particularly in education, where students might rely on AI for their homework. The company’s ChatGPT app gained significant popularity, sparking alarm about accuracy, safety, and cheating. Additionally, misinformation through AI-generated text has become a growing concern, as AI-generated tweets are more convincing than those written by humans. Governments have not yet established regulations for AI and are leaving it to individual groups and organizations to develop their own protective measures. OpenAI has not provided further comment on the matter.
The Turing test, proposed by Alan Turing in 1950, has long been considered a benchmark for testing AI. However, the test has been criticized for being too vague and focused on deception rather than genuine intelligence. The rise of large language models (LLMs) like ChatGPT has renewed interest in the role of language in evaluating and creating AI. These models, purely based on language, have impressive capabilities, including conversation, essay writing, and coding. However, the underlying mechanisms behind their behavior are still not well understood. Scientists are working to uncover the capabilities and mechanisms of LLMs, often compared to “alien intelligence.” Understanding LLMs is crucial for their application in medicine and law and requires new, more systematic tests. While LLMs excel at tasks like exams, they can be brittle and unreliable in real-world scenarios. Researchers are divided on whether LLMs demonstrate reasoning and understanding or if their limited reliability suggests otherwise. More transparent disclosures from AI companies and increased scrutiny from regulatory bodies could help shed light on LLMs and their inner workings.