Back to home

Compare

Comparing: 'Did AI make this?' — Anthropic CEO's blunt take on trust & 客户问你这活是不是 AI 做的 — 看看 Anthropic 老总怎么回

AEN
AnthropicAI trustclient communication·

'Did AI make this?' — Anthropic CEO's blunt take on trust

Last week, Xiaomei's client saw through it in 30 seconds

My designer friend Xiaomei delivered a private logo gig last week. The client looked up 30 seconds after opening the file: "This was generated by AI, right?" Xiaomei froze for 3 seconds, then squeezed out "It's assisted..." The room went cold. I felt my stomach drop hearing this — I'd been in that exact spot before, and it took me a while to recover.

Anthropic's CEO is wrestling with the same thing

Anthropic CEO Dario Amodei broke his silence on X last week to respond to critics. He said "we're safe" PR alone won't rebuild trust — the only way is to "actually cure cancer," letting users see real, tangible results. He also addressed the question "is AI safety talk backfiring," admitting it's a real problem.

Translated: clients don't care what tools you used behind the scenes, they care whether the deliverable is useful. Xiaomei slipped up by dodging the question. If she'd said "I used AI to draft 3 directions, then I manually color-graded this final version," it would have scored her points.

One step you can try today

  • Money: $0
  • Time: 30 minutes
  • Technical barrier: none — if you can write, you can do this
  • First step: grab a piece of paper, write down 3 things you used AI to help with this week, and after each one write "what the client got out of it"

Not for posting on social media — this is for you to know in your own head. Next time someone asks, you can answer in 3 seconds flat.

Pick based on where you are

  • Just starting, few clients: no need to publicly say "I use AI" right now. Make the work good first; tools come later.
  • 1–2 stable clients: proactively mention "I compared 3 directions for this version" on delivery — putting the process on the table feels more solid than hiding it.
  • Scaling up, signing long-term contracts: write your AI usage into your service description, e.g., "AI-assisted drafts, human-vetted finals." Saying it upfront saves you from explaining later.

I also got stuck on the "do I admit it when asked" hurdle. This approach isn't for everyone, and it's fine to skip for now. But next time a client opens their mouth to ask, at least don't freeze for those 3 seconds again.

BZH
AnthropicAI信任客户沟通·

客户问你这活是不是 AI 做的 — 看看 Anthropic 老总怎么回

上周小美的客户,30 秒就识破了

我做设计的朋友小美,上周交一个 logo 私单。客户打开文件 30 秒后抬头问:"这是 AI 跑出来的吧?"小美愣了 3 秒,憋出一句"是辅助的……"气氛当场冷掉。我听完心里咯噔一下——这场景我之前也踩过,差点没缓过来。

Anthropic 老总最近也在纠结同一件事

Anthropic 的 CEO Dario Amodei 上周罕见地在 X 上回应批评。他说,光靠"我们很安全"的宣传挽回不了信任,唯一办法是"真正把癌症治好"——让用户看到实打实的结果。他也回应了"AI 安全言论是不是在帮倒忙"的质疑,承认这是个真问题。 翻译一下:客户不在乎你背后用了什么工具,在乎的是交付的东西有没有用。小美那次错在回避问题;如果她当时说"我用 AI 提了 3 稿方向,最后这版是我手动调色的",反而加分。

你今天能试的一步

  • 钱:0 元
  • 时间:30 分钟
  • 技术门槛:不用懂任何技术,会写字就行
  • 第一步:拿张纸,写下你这周用 AI 帮过的 3 件事,每件事后面写一句"客户得到了什么"
不是让你发朋友圈,是让你自己心里有数。下次被问到,你能 3 秒内接住。

按你的情况挑

  • 刚起步、没几个客户:现在不用急着公开说"我用 AI",先把活做漂亮,工具是后面的事。
  • 有 1-2 个稳定客户:交付时主动说"这版我比了 3 个方向",把过程感摆出来,比藏着踏实。
  • 在扩规模、要签长期合同:把 AI 用法写进你的服务说明,比如"初稿 AI 辅助,终稿人工把关",提前讲清楚比事后解释省力。
我也卡过"被问要不要承认"这个坎。这思路不是所有人都需要,现在不试也没事。但下次客户开口问你,至少别再愣那 3 秒。