California-based start-up OpenAI Inc has released a chatbot capable of answering a variety of questions, but its impressive performance has reopened the debate on the risks linked to artificial intelligence (AI) technologies.
The conversations with ChatGPT, posted on Twitter by fascinated users, show a kind of omniscient machine, capable of explaining scientific concepts and writing scenes for a play, university dissertations or even functional lines of computer code.
“Its answer to the question ‘what to do if someone has a heart attack’ was incredibly clear and relevant,” said Claude de Loupy, head of Syllabs, a French company specialized in automatic text generation.
Photo: AFP
“When you start asking very specific questions, ChatGPT’s response can be off the mark,” but its overall performance remains “really impressive,” with a “high linguistic level,” he said.
OpenAI, cofounded in 2015 in San Francisco by Tesla Inc CEO Elon Musk, who left the AI business in 2018, received US$1 billion from Microsoft Corp in 2019.
The start-up is best known for its automated creation software: GPT-3 for text generation and DALL-E for image generation.
ChatGPT is able to ask its interlocutor for details, and has fewer strange responses than GPT-3, which, in spite of its prowess, sometimes spits out absurd results, De Loupy said.
“A few years ago chatbots had the vocabulary of a dictionary and the memory of a goldfish,” said Sean McGregor, a researcher who runs a database of AI-related incidents.
“Chatbots are getting much better at the ‘history problem,’ where they act in a manner consistent with the history of queries and responses,” McGregor said. “The chatbots have graduated from goldfish status.”
Like other programs relying on deep learning, mimicking neural activity, ChatGPT has one major weakness: “It does not have access to meaning,” De Loupy said.
The software cannot justify its choices, such as explaining why it picked the words that make up its responses.
However, AI technologies that are able to communicate are increasingly able to give an impression of thought.
Researchers at Facebook parent Meta Platforms Inc recently developed a computer program dubbed Cicero, after the Roman statesman.
The software has proven proficient at the board game Diplomacy, which requires negotiation skills.
“If it doesn’t talk like a real person — showing empathy, building relationships and speaking knowledgeably about the game — it won’t find other players willing to work with it,” Meta said in a report.
In October, Character.ai, a start-up founded by former Google engineers, put an experimental chatbot online that can adopt any personality.
Users can create characters based on a brief description and “chat” with a fake Sherlock Holmes, Socrates or former US president Donald Trump.
This level of sophistication fascinates, but also worries some observers, who voice concern these technologies could be misused to trick people by spreading false information or by creating increasingly credible scams.
What does ChatGPT think of these hazards?
“There are potential dangers in building highly sophisticated chatbots, particularly if they are designed to be indistinguishable from humans in their language and behavior,” the chatbot told Agence France-Presse.
Some businesses are putting safeguards in place to avoid abuse of their technologies.
On its welcome page, OpenAI has a disclaimer saying the chatbot “may occasionally generate incorrect information” or “produce harmful instructions or biased content.”
SEMICONDUCTOR SERVICES: A company executive said that Taiwanese firms must think about how to participate in global supply chains and lift their competitiveness Taiwan Semiconductor Manufacturing Co (TSMC, 台積電) yesterday said it expects to launch its first multifunctional service center in Pingtung County in the middle of 2027, in a bid to foster a resilient high-tech facility construction ecosystem. TSMC broached the idea of creating a center two or three years ago when it started building new manufacturing capacity in the US and Japan, the company said. The center, dubbed an “ecosystem park,” would assist local manufacturing facility construction partners to upgrade their capabilities and secure more deals from other global chipmakers such as Intel Corp, Micron Technology Inc and Infineon Technologies AG, TSMC said. It
People walk past advertising for a Syensqo chip at the Semicon Taiwan exhibition in Taipei yesterday.
NO BREAKTHROUGH? More substantial ‘deliverables,’ such as tariff reductions, would likely be saved for a meeting between Trump and Xi later this year, a trade expert said China launched two probes targeting the US semiconductor sector on Saturday ahead of talks between the two nations in Spain this week on trade, national security and the ownership of social media platform TikTok. China’s Ministry of Commerce announced an anti-dumping investigation into certain analog integrated circuits (ICs) imported from the US. The investigation is to target some commodity interface ICs and gate driver ICs, which are commonly made by US companies such as Texas Instruments Inc and ON Semiconductor Corp. The ministry also announced an anti-discrimination probe into US measures against China’s chip sector. US measures such as export curbs and tariffs
The US on Friday penalized two Chinese firms that acquired US chipmaking equipment for China’s top chipmaker, Semiconductor Manufacturing International Corp (SMIC, 中芯國際), including them among 32 entities that were added to the US Department of Commerce’s restricted trade list, a US government posting showed. Twenty-three of the 32 are in China. GMC Semiconductor Technology (Wuxi) Co (吉姆西半導體科技) and Jicun Semiconductor Technology (Shanghai) Co (吉存半導體科技) were placed on the list, formally known as the Entity List, for acquiring equipment for SMIC Northern Integrated Circuit Manufacturing (Beijing) Corp (中芯北方積體電路) and Semiconductor Manufacturing International (Beijing) Corp (中芯北京), the US Federal Register posting said. The
RSS