Can I randomise the scenario played in the OpenAI Gym Atari environments?

Question

I am looking at the Seaquest Atari game in OpenAI Gym and I have the following code snippet:

import gym
import time
import random

env = gym.make('Seaquest-v4')

for i_episode in range(10):
    seed = (int)(time.time())    
    random.seed(seed)
    env.seed(seed)
    observation = env.reset()
    for t in range(800):
        env.render()
        time.sleep(0.01)
        action = 0
        observation, reward, done, info = env.step(action)
        if done:
            print("Episode finished after {} timesteps".format(t+1))
            break
env.close()

The code above renders 10 games, however, even though I reset the seed every time to be unique, I still get the same scenario of the game. As a result the agent is trained on the same scenario. Could someone suggest a way to fix this? I am just trying to find a way to make the Atari emulator to produce different scenarios of the game.