AT&T, T-Mobile tighten AI guardrails as autonomous networks remain years away

FILE PHOTO: The AT&T logo is seen in a store window, as airports around the country are awaiting for Verizon and AT&T to rollout their 5G technology, in the Manhattan borough of New York City, New York, U.S., January 19, 2022. REUTERS/Brendan McDermid/File Photo

Oct 23 (Reuters) - Telecom giants AT&T and T-Mobile are investing to safeguard customer data and network ⁠infrastructure from new cybersecurity threats as AI integration ramps up, though fully autonomous networks ‌remain years away, company executives told Reuters.

As AI agents automateworkflows anddata analysis, with declining human intervention,cybersecurity concerns are growing across industries over access to sensitive information and connected software.

For telecom operators, which manage vast amounts of customer data and critical communications infrastructure, ​vulnerabilities in AI systems or third-party software could extend from data ⁠breaches to network disruptions.

"Yes, we are increasing ⁠our investments in how we monitor and react," T-Mobile's Chief Technology Officer John Saw told the Reuters ⁠Global ‌Markets Forum. The company simultaneously works with vendors to ensure security holes are closed as soon as possible, he added.

The Bellevue, Washington-based wireless carrier has suffered several data breaches, including incidents ⁠in 2021 and 2023, cumulatively impacting over 100 million customer accounts. ​It hasagreed to various settlements ‌with consumers and the Federal Communications Commission involving cybersecurity penalties and investments.

AT&T, meanwhile, warned that ⁠third-party software providers ​and AI agents pose cybersecurity risks that extend beyond operators' own networks.

"There (are) industry problems, not things that AT&T can solve by itself," CTO Jeremy Legg said, questioning if the open-source community followed comparable security standards.

The Texas-based company has ⁠deployed AI governance controls such as encryption,multi-factor authentication and human ​oversight of data access. It is also working to verify whether third-party AI agents seeking access to customer accounts are acting on behalf of legitimate customers, the company said.

AI has automated aspects of network management for ⁠telecom operators, but handing over entire networks to AI-driven autonomous systems is still "years away," Legg said, as human intervention is still required to solve complex network problems.

For instance, AT&T has automated some tasks such as analyzing trouble tickets to identify potential fixes and, in some cases, automatically implementing repairs. T-Mobile, meanwhile, uses ​AI to automatically adjust cell-site antennas when nearby sites go down to ⁠compensate for coverage gaps.

During Winter Storm Fern in January, T-Mobile's self-organizing network, or SON, restored coverage to ​about 66% of customers within an hour and 90% within eight ‌hours, with most customers regaining service before power was ​restored, Saw said.

(Join GMF, a chat room hosted on LSEG Messenger for live interviews: )

(Reporting by Mehnaz Yasmin and Harshita Mary Varghese in Bengaluru; Editing by Divya Chowdhury and Joyjeet Das)

Thank you for your report!

SkyeChip collaborates with Japan's SAIMEMORY on next-generation AI memory technology

Proposed export curbs target advanced solar technologies

Soccer-Police to deploy live facial recognition technology at Chelsea-Spurs game

资料照片：2022年1月19日，美国纽约州纽约市曼哈顿区，一家商店橱窗里可以看到AT&T的标志。全国各地的机场都在等待Verizon和AT&T推出其5G技术。（路透社/Brendan McDermid/资料照片）

10 月 23 日（路透社）——电信巨头 AT&T 和 T-Mobile 的高管告诉路透社，随着人工智能集成力度的加大，这两家公司正在投资保护客户数据和网络基础设施免受新的网络安全威胁，尽管完全自主的网络距离实现还有数年时间。

随着人工智能代理实现工作流程和数据分析的自动化，人类干预减少，各行各业对敏感信息和联网软件的访问安全问题日益关注。

对于管理大量客户数据和关键通信基础设施的电信运营商而言，人工智能系统或第三方软件中的漏洞可能导致从数据泄露到网络中断等一系列问题。

T-Mobile首席技术官约翰·索在路透全球市场论坛上表示：“是的，我们正在加大对监控和应对方式的投资。”他还补充说，公司同时也在与供应商合作，确保尽快修复安全漏洞。

这家总部位于华盛顿州贝尔维尤的无线运营商遭遇了多次数据泄露事件，包括2021年和2023年的事件，累计影响超过1亿个客户账户。该公司已与消费者和联邦通信委员会达成多项和解协议，涉及网络安全罚款和投资。

与此同时，AT&T 警告称，第三方软件提供商和人工智能代理带来的网络安全风险超出了运营商自身的网络范围。

“这是行业问题，不是 AT&T 自己能解决的问题，”首席技术官 Jeremy Legg 说，他质疑开源社区是否遵循类似的安全性标准。

这家总部位于德克萨斯州的公司已部署了人工智能治理控制措施，例如加密、多因素身份验证和人工数据访问监督。该公司表示，他们还在努力验证试图访问客户帐户的第三方人工智能代理是否代表合法客户行事。

莱格表示，人工智能已经实现了电信运营商网络管理的某些方面的自动化，但将整个网络移交给人工智能驱动的自主系统仍然“遥遥无期”，因为解决复杂的网络问题仍然需要人工干预。

例如，AT&T 已将一些任务自动化，例如分析故障单以确定潜在的修复方案，并在某些情况下自动执行修复。与此同时，T-Mobile 利用人工智能技术，在附近基站出现故障时自动调整基站天线，以弥补覆盖盲区。

索表示，在 1 月份的冬季风暴 Fern 期间，T-Mobile 的自组织网络 (SON) 在一小时内恢复了约 66% 用户的网络覆盖，并在八小时内恢复了 90% 用户的网络覆盖，大多数用户在电力恢复之前就恢复了服务。

（加入GMF，这是一个在LSEG Messenger上举办的聊天室，可进行实时访谈：）

（Mehnaz Yasmin 和 Harshita Mary Varghese 在班加罗尔报道；Divya Chowdhury 和 Joyjeet Das 编辑）

感谢您的报告！

SkyeChip与日本SAIMEMORY合作开发下一代AI存储技术

拟议的出口限制措施针对先进太阳能技术

足球警察将在切尔西对阵热刺的比赛中部署实时人脸识别技术