去年，隨著人工智能 (AI) 成為日常使用中更加重要的工具，圍繞該技術(shù)的法律環(huán)境也開始發(fā)展。

從開始形成的全球法規(guī)和法律到無數(shù)指控版權(quán)和數(shù)據(jù)侵權(quán)的訴訟，人工智能受到了每個人的關(guān)注。

隨著 2024 年的臨近，Cointelegraph 請求從事法律和人工智能交叉領(lǐng)域工作的業(yè)內(nèi)人士幫助分析 2023 年的教訓(xùn)以及它們對未來一年的意義。

要全面了解 2023 年人工智能領(lǐng)域發(fā)生的事情，請不要忘記查看 Cointelegraph 的“2023 年人工智能終極指南”。

歐盟人工智能法案執(zhí)行延遲

2023年，歐盟成為首批在通過立法以規(guī)范高級人工智能模型的部署和開發(fā)方面取得重大進(jìn)展的地區(qū)之一。

《歐盟人工智能法案》最初于4月提出，并于6月獲得議會通過。

12月8日，歐洲議會和理事會談判代表就該法案達(dá)成臨時協(xié)議。

一旦完全生效，它將規(guī)范政府在生物識別監(jiān)控中使用人工智能，監(jiān)督 ChatGPT 等大型人工智能系統(tǒng)，并制定開發(fā)人員在進(jìn)入市場之前應(yīng)遵循的透明度規(guī)則。

然而，該法案已經(jīng)受到科技行業(yè)“過度監(jiān)管”的批評。

由于開發(fā)商的抵制和延誤記錄，貝克·麥堅(jiān)時 (Baker McKenzie) 合伙人、《

的作者 Lothar Determann告訴 Cointelegraph：

“隨著歐盟人工智能法案的頒布，我們可能會看到類似的延遲時間表，這似乎并非完全不可能?！?/p>

德特曼指出，雖然該協(xié)議已于12月初達(dá)成，但最終文本仍有待觀察。

他補(bǔ)充說，包括法國總統(tǒng)在內(nèi)的幾位主要成員國的政治家都對當(dāng)前的草案表示了擔(dān)憂。

“這讓我想起了電子隱私法規(guī)的軌跡，歐盟于 2016 年宣布該法規(guī)將于 2018 年 5 月與《通用數(shù)據(jù)保護(hù)法規(guī)》一起生效，但五年后仍未最終確定。”

Laura De Boel, a partner in the Brussels office of law firm Wilson Sonsini Goodrich & Rosati, also pointed out that the December development is a “political agreement,” with formal adoption yet to come in early 2024.

She explained further that EU lawmakers have included a “phased grace period,” during which:

“The rules on prohibited AI systems will apply after six months, and the rules on General Purpose AI will apply after 12 months,” she said. “The other requirements of the AI Act will apply after 24 months, except that the obligations for high-risk systems defined in Annex II will apply after 36 months.”

Compliance challenges

Despite a flurry of new regulations entering the scene, 2024 will present some challenges for companies in terms of compliance.

De Boel said that the European Commission has already called on AI developers to voluntarily implement the key obligations of the AI Act even before they become mandatory:

“They will need to start building the necessary internal processes and prepare their staff.”

However, Determann said that even without a comprehensive AI regulatory scheme, “we’ll see compliance challenges as businesses grapple with the application of existing regulatory schemes to AI.”

This includes the EU General Data Protection Regulation (GDPR), privacy laws around the world, intellectual property laws, product safety regulations, property laws, trade secrets, confidentiality agreements and industry standards, among others.

To this note, in the United States, the administration of President Joe Biden issued a lengthy executive order on Oct. 30 intended to protect citizens, government agencies and companies by ensuring AI safety standards.

The order established six new standards for AI safety and security, including intentions for ethical AI usage within government agencies.

While Biden is quoted saying that the order aligns with the government’s principles of “safety, security, trust, openness,” insiders in the industry said it has created a “challenging” climate for developers.

This primarily boils down to discerning concrete compliance standards out of vague language.

In a previous interview with Cointelegraph, Adam Struck, a founding partner at Struck Capital and an AI investor, told Cointelegraph that the order makes it tricky for developers to anticipate future risks and compliance according to the legislation, which is based on assumptions about products that aren’t fully developed yet. He said:

“This is certainly challenging for companies and developers, particularly in the open-source community, where the executive order was less directive.”

Related: ChatGPT’s first year marked by existential fear, lawsuits and boardroom drama

More specific laws

Another anticipation in the legal landscape of 2024 is more specific, narrowly framed laws. This can already be seen as some countries deploy regulations against AI-generated deepfakes.

Regulators in the U.S. are already considering introducing regulations on political deepfakes in the lead-up to the 2024 presidential elections. As of late November, India has begun finalizing laws against deepfakes.

Determann cautioned AI-related businesses and those using AI products:

“Moving forward, businesses will need to stay up-to-date on these developments, which will include disclosure requirements for bots, restrictions on ‘deepfakes’ and audit requirements for job application evaluation systems.”

He continued to say that such specifically focused laws tend to have a “better chance” to result in the intended impact than overly broad regulations.

“This is because businesses can understand and comply with them more easily, and authorities can enforce them more effectively,” he explained.

There have also been rumblings that 2024 could see regulations focusing on investments in technology. Among lawmakers in the U.S., there has already been talk of regulating investments.

In July, the U.S. Senate backed bipartisan legislation that requires local companies to report any investments in Chinese technologies, particularly semiconductors used in AI.

Copyright clarification

Many in the industry also expect to see a clarification of what is deemed copyright infringement and violation of copyright law.

Over the last year, there have been a number of high-profile copyright-related lawsuits accusing leading AI developers like OpenAI, Microsoft, Meta and Google of abusing copyright laws.

These lawsuits have covered almost every form of content, from art and music, to literature and news.

Most recently, on Dec. 7, The New York Times filed a case against OpenAI and Microsoft for copyright infringement of their news content during the training of AI models. Prominent lawyers in the space have called this particular case a potential “watershed moment” for the AI and copyright space.

With the majority of these cases ongoing, 2024 is expected to bring some sort of resolution to this issue — whether wholly or partially remains to be seen.

What does AI have to say?

In the spirit of all things AI, Cointelegraph decided to ask ChatGPT itself what it believes will come of the legal scene surrounding AI in the upcoming year.

When asked, “Can you give me some predictions for AI law in 2024?” ChatGPT responded with the following assumptions:

• More “strict regulations” globally about ethical concerns and biases.
• Developing clearer “l(fā)iability standards” for AI-related incidents for developers and users.
• A heightened “emphasis on data privacy laws.”
• An “increased demand for transparency in AI algorithms.”
• The emergence of “specialized boards or regulatory bodies” dedicated to overseeing AI implementation.
• Potential updates to labor laws and new employment models with AI in consideration.
• Efforts to establish a “cohesive international framework for governing AI technologies.”
• Implementation of “measures to protect consumers” from misleading or harmful AI applications.

ChatGPT 是否有所作為，只有時間才能證明，2024 年將是告訴我們所有人的一年。

請務(wù)必在 2024 年關(guān)注此空間，了解有關(guān)人工智能的最新更新。

雜志：

熱點(diǎn)：人工智能 卡爾 挑戰(zhàn) 法律