到目前為止，世界上大多數(shù)企業(yè)已經(jīng)意識(shí)到生成式人工智能將繼續(xù)存在，他們應(yīng)該至少在一定程度上使用它來(lái)進(jìn)行運(yùn)營(yíng)。

但如何以安全可靠的方式做到這一點(diǎn)，并實(shí)現(xiàn)“人工智能治理”（該術(shù)語(yǔ)用于跟蹤組織內(nèi)的人工智能部署和數(shù)據(jù)使用情況）？

總部位于紐約市和比利時(shí)布魯塞爾的數(shù)據(jù)治理公司 Collibra 正在推出專(zhuān)門(mén)為解決這一問(wèn)題而量身定制的新解決方案。

Collibra 聯(lián)合創(chuàng)始人兼首席執(zhí)行官 Felix Van de Maele 在獨(dú)家視頻通話采訪中表示：“人們一致認(rèn)為，人工智能正在成為幾乎每個(gè)組織的關(guān)注焦點(diǎn)，但每個(gè)組織都明白他們需要以正確的方式開(kāi)發(fā)人工智能?！迸c VentureBeat 合作。

今天發(fā)布的新產(chǎn)品 Collibra AI Governance 有望通過(guò)確保部署的安全性、效率和可信性來(lái)重新定義組織實(shí)施 AI 的方式。

VB事件

AI 影響之旅 – 紐約

我們將于 2 月 29 日在紐約與微軟合作，討論如何平衡人工智能應(yīng)用的風(fēng)險(xiǎn)和回報(bào)。

請(qǐng)求邀請(qǐng)參加以下獨(dú)家活動(dòng)。

請(qǐng)求邀請(qǐng)

彌合組織技術(shù)和政策之間的差距

它作為 Collibra 數(shù)據(jù)智能平臺(tái)的一個(gè)組成部分而開(kāi)發(fā)，旨在彌合數(shù)據(jù)、人工智能和法律合規(guī)性之間的關(guān)鍵差距。

它使這些領(lǐng)域的團(tuán)隊(duì)能夠有效協(xié)作，確保人工智能項(xiàng)目符合法律和隱私要求，最大限度地降低數(shù)據(jù)風(fēng)險(xiǎn)，并提高模型性能和投資回報(bào) (ROI)，從而加快從概念到生產(chǎn)的過(guò)程。

Van de Maele 告訴 VentureBeat，它還通過(guò)在整個(gè)人工智能采用過(guò)程中的每一步為組織提供幫助，從“構(gòu)思到交易到模型開(kāi)發(fā)，到實(shí)際投入生產(chǎn)并持續(xù)監(jiān)控”。

“這一切都是關(guān)于‘如何在整個(gè)生命周期中支持整個(gè)人工智能用例？’”

這不僅適用于一種人工智能解決方案或平臺(tái)，還適用于任何給定組織使用人工智能的所有方式。

Collibra AI Governance 的獨(dú)特之處在于提供對(duì) AI 項(xiàng)目的直觀管理和監(jiān)控，確保敏感數(shù)據(jù)保護(hù)，并提高對(duì) AI 用例的透明度和理解。

“The biggest challenge is ensuring true visibility across the whole organization about who’s doing what and how exactly,” Van de Maele said. “It’s more than documentation. It’s also the workflow...to ensure that things are done in the right way.”

It achieves this by sitting atop and integrating with existing Collibra data management offerings for customers, including the Collibra Data Catalog, Collibra Data Quality & Observability, Collibra Data Privacy, and Collibra Protect.

A new approach to risk management and mitigation

The imperative for robust AI governance is underscored by findings from a recent IDC survey, which revealed that while 71% of companies are already harnessing AI, the pathway to its successful implementation is fraught with potential pitfalls.

From model bias and inaccuracies to legal and ethical conundrums, the challenges are many, posing significant risks to an organization’s reputation and financial health. Just take a look at how Google’s reputation and even its stock valuation have suffered in the days since users began criticizing its Gemini AI model for producing racially confused imagery and philosophically bizarre responses.

Collibra’s AI Governance solution is meticulously crafted to mitigate these risks by providing a comprehensive framework for AI governance.

This includes streamlined compliance processes, risk mitigation strategies, and mechanisms to enhance model performance and ROI. By enabling a unified platform for governing both AI and the underlying data, Collibra aims to dispel the concerns surrounding AI adoption.

And all of this is included as an “add-on” module to Collibra’s existing Data Intelligence Platform.

Laura Sellers, Collibra’s chief product officer, provided a statement in a press release that emphasized the product’s value proposition, noting the prevalent organizational desire to prioritize AI amidst concerns over its unpredictability and the potential for adverse outcomes.

“With Collibra AI Governance, we’re helping to eliminate the concerns of CIOs, CDOs, data leaders, and legal teams by providing a unified platform to govern AI and the data that drives it with automated workflows, processes, and policies to deliver trusted AI across the organization,” Sellers stated.

Endorsements from key figures in the wider enterprise software world

Industry leaders are further validating and endorsing Collibra’s new offering. Dan Yu, Senior Vice President of Marketing and Solutions, Data and Analytics at SAP SE, lauded Collibra AI Governance for its potential to ensure the ethical and responsible use of AI across organizations.

As Collibra AI Governance is set to be showcased at the upcoming Gartner Data & Analytics Summit in Orlando, Florida, from March 11-13, 2024.

By addressing the complex challenges that come with AI projects, Collibra is not just offering a product but a promise of AI that is both powerful and principled.